Introduction And Methodology
Working in statistics, I’m familiar with the Gartner Hype Cycle. for those unfamiliar, it’s a way of identifying “the maturity, adoption and social application of specific technologies”. I won’t comment about their methodology or validity- because I don’t know it- but I was interested in taking this idea and applying it to cats, so I did. (Disclaimer: I’m a dog person.)
The data I had available was the Google Trends API. I wanted to see search volume over time for each breed of cat. I could use this as a time series to do my own Carmen Hype Cycle Analysis for cats. The trend data that you get back from Google is relative- I’ll quote directly from Google, “Numbers represent search interest relative to the highest point on the chart for the given region and time. A value of 100 is the peak popularity for the term. A value of 50 means that the term is half as popular. Likewise a score of 0 means the term was less than 1% as popular as the peak”.
To get around this I included a baseline term so I could scale the results to something that could be compared across searches. The next step was to get a list of cat breeds, which I learned is actually a contested number. I chose to scrape Wikipedia for the list of cat breeds. I could have copied and pasted the cat breeds, but I wanted to work on my HTML parsing, and R has a fantastic rvest package that does a wonderful job.
Results
Here is an example of the normalized results from the Google API. It is the popularity of cat breeds over time.
It’s a nice graph, but it only shows 5 cat breeds, and in total I have 96. Way too many to view with this visualization. Instead we can collapse the time aspect and plot the trend and popularity as X and Y. The popularity is the mean search results for the time span and the trend is the search results divided by a moving average. This gives us the following plot.
Looking at the plot, and based on the newly patented Carmen Hype Cycle for Cats, I’m predicting the German Rex Cat to be an up-and-comer for 2018. Meanwhile, it seems people have become disenchanted with the oriental longhair.
In reality, I used this analysis to work with the Google API, scraping HTML tables form Wikipedia, and some nice HTML interactive graphs.